Closed mdedetrich closed 1 week ago
Reading the Akka issue and PR seems to indicate that this will cause test failures. I can help if needed to try to fix up the test issues. I'm hoping to get Pekko HTTP 1.1.0 released but it seems useful to fix this.
Reading the Akka issue and PR seems to indicate that this will cause test failures. I can help if needed to try to fix up the test issues. I'm hoping to get Pekko HTTP 1.1.0 released but it seems useful to fix this.
The test failures are expected because some of the tests expect the stream to be cancelled within 100ms and now that is has been increased to 1000ms those tests are now broken (as per @jrudolph 's comment, the change itself is safe).
Ill have time to look into this tonight
Should this actually be marked as closing the issue? It's better than the current status quo, but the underlying issue is still there.
For reference, we've used 1s
for a long time with no issue until today, where we saw this resurface again. So 1s
is not bullet proof.
Should this actually be marked as closing the issue? It's better than the current status quo, but the underlying issue is still there.
For reference, we've used
1s
for a long time with no issue until today, where we saw this resurface again. So1s
is not bullet proof.
Fair point, ill change this so it won't close original issue
PR is ready, I managed to fix the issues with the tests.
@pjfanning Feel free to merge when you see fit
References: https://github.com/apache/pekko-http/issues/422
Still need to update tests as per @jrudolph original comment, i.e.